[CI] Try introducing isort. (#3495)

vllm/model_executor/layers/vocab_parallel_embedding.py

@@ -4,13 +4,11 @@ import torch
 import torch.nn.functional as F
 from torch.nn.parameter import Parameter
 
-from vllm.model_executor.parallel_utils.parallel_state import (
-    get_tensor_model_parallel_rank,
-    get_tensor_model_parallel_world_size,
-)
-from vllm.model_executor.parallel_utils.utils import divide
 from vllm.model_executor.parallel_utils.communication_op import (
     tensor_model_parallel_all_reduce)
+from vllm.model_executor.parallel_utils.parallel_state import (
+    get_tensor_model_parallel_rank, get_tensor_model_parallel_world_size)
+from vllm.model_executor.parallel_utils.utils import divide
 from vllm.model_executor.utils import set_weight_attrs
 
 DEFAULT_VOCAB_PADDING_SIZE = 64

vllm/model_executor/models/baichuan.py

@@ -32,11 +32,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_rank, get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/chatglm.py

@@ -15,11 +15,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/deepseek.py

@@ -33,14 +33,14 @@ from vllm.model_executor.layers.fused_moe import fused_moe
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
-                                               ReplicatedLinear,
                                                QKVParallelLinear,
+                                               ReplicatedLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.communication_op import (
     tensor_model_parallel_all_reduce)
 from vllm.model_executor.parallel_utils.parallel_state import (

vllm/model_executor/models/falcon.py

@@ -32,8 +32,8 @@ from vllm.model_executor.layers.linear import (ColumnParallelLinear,
                                                LinearMethodBase,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
     VocabParallelEmbedding)

vllm/model_executor/models/gemma.py

@@ -28,8 +28,8 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
     VocabParallelEmbedding)

vllm/model_executor/models/gpt_j.py

@@ -28,11 +28,11 @@ from vllm.model_executor.layers.linear import (ColumnParallelLinear,
                                                LinearMethodBase,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/gpt_neox.py

@@ -28,11 +28,11 @@ from vllm.model_executor.layers.linear import (ColumnParallelLinear,
                                                LinearMethodBase,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/internlm2.py

@@ -12,11 +12,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/jais.py

@@ -24,29 +24,23 @@ from typing import List, Optional
 
 import torch
 from torch import nn
-from vllm.transformers_utils.configs import JAISConfig
 
 from vllm.attention import Attention, AttentionMetadata
-from vllm.model_executor.layers.linear import (
-    ColumnParallelLinear,
-    LinearMethodBase,
-    QKVParallelLinear,
-    RowParallelLinear,
-)
+from vllm.model_executor.layers.linear import (ColumnParallelLinear,
+                                               LinearMethodBase,
+                                               QKVParallelLinear,
+                                               RowParallelLinear)
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, )
+    VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
-    get_tensor_model_parallel_world_size,
-    get_tensor_model_parallel_rank,
-)
-from vllm.model_executor.weight_utils import (
-    default_weight_loader,
-    hf_model_weights_iterator,
-)
-from vllm.sequence import SamplerOutput
+    get_tensor_model_parallel_rank, get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata
+from vllm.model_executor.weight_utils import (default_weight_loader,
+                                              hf_model_weights_iterator)
+from vllm.sequence import SamplerOutput
+from vllm.transformers_utils.configs import JAISConfig
 
 
 class SwiGLUActivation(nn.Module):

vllm/model_executor/models/llama.py

@@ -35,11 +35,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead, DEFAULT_VOCAB_PADDING_SIZE)
+    DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/mixtral.py

@@ -35,11 +35,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                QKVParallelLinear,
                                                ReplicatedLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead, DEFAULT_VOCAB_PADDING_SIZE)
+    DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.communication_op import (
     tensor_model_parallel_all_reduce)
 from vllm.model_executor.parallel_utils.parallel_state import (

vllm/model_executor/models/mixtral_quant.py

@@ -24,24 +24,22 @@
 from typing import List, Optional
 
 import numpy as np
-
 import torch
 import torch.nn.functional as F
-
 from torch import nn
 from transformers import MixtralConfig
 
 from vllm.attention import Attention, AttentionMetadata
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (LinearMethodBase,
-                                               ReplicatedLinear,
                                                QKVParallelLinear,
+                                               ReplicatedLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.communication_op import (
     tensor_model_parallel_all_reduce)
 from vllm.model_executor.parallel_utils.parallel_state import (

vllm/model_executor/models/olmo.py

@@ -40,32 +40,27 @@ from typing import List, Optional, Tuple
 
 import torch
 import torch.nn.functional as F
+# this model must need this dependency
+from hf_olmo import OLMoConfig
 from torch import nn
 
 from vllm.attention import Attention, AttentionMetadata
-from vllm.model_executor.layers.linear import (
-    ColumnParallelLinear,
-    LinearMethodBase,
-    QKVParallelLinear,
-    RowParallelLinear,
-)
-from vllm.model_executor.layers.rotary_embedding import get_rope
+from vllm.model_executor.layers.linear import (ColumnParallelLinear,
+                                               LinearMethodBase,
+                                               QKVParallelLinear,
+                                               RowParallelLinear)
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
     VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
-    get_tensor_model_parallel_world_size, )
+    get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata
-from vllm.model_executor.weight_utils import (
-    default_weight_loader,
-    hf_model_weights_iterator,
-)
+from vllm.model_executor.weight_utils import (default_weight_loader,
+                                              hf_model_weights_iterator)
 from vllm.sequence import SamplerOutput
 
-# this model must need this dependency
-from hf_olmo import OLMoConfig
-
 
 class SwiGLU(nn.Module):
 

vllm/model_executor/models/orion.py

@@ -16,11 +16,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/phi.py

@@ -47,11 +47,11 @@ from vllm.model_executor.layers.linear import (ColumnParallelLinear,
                                                LinearMethodBase,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/qwen.py

@@ -17,11 +17,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/qwen2.py

@@ -29,24 +29,24 @@ from torch import nn
 from transformers import Qwen2Config
 
 from vllm.attention import Attention, AttentionMetadata
+from vllm.config import LoRAConfig
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata
 from vllm.model_executor.weight_utils import (default_weight_loader,
                                               hf_model_weights_iterator)
 from vllm.sequence import SamplerOutput
-from vllm.config import LoRAConfig
 
 
 class Qwen2MLP(nn.Module):

vllm/model_executor/models/stablelm.py

@@ -31,11 +31,11 @@ from vllm.model_executor.layers.linear import (LinearMethodBase,
                                                MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead)
+    ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
 from vllm.model_executor.sampling_metadata import SamplingMetadata

vllm/model_executor/models/starcoder2.py

@@ -25,19 +25,19 @@ from torch import nn
 from transformers import Starcoder2Config
 
 from vllm.attention import Attention, AttentionMetadata
-from vllm.model_executor.sampling_metadata import SamplingMetadata
 from vllm.model_executor.layers.activation import get_act_fn
-from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.linear import (ColumnParallelLinear,
                                                LinearMethodBase,
                                                QKVParallelLinear,
                                                RowParallelLinear)
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.rotary_embedding import get_rope
 from vllm.model_executor.layers.sampler import Sampler
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    VocabParallelEmbedding, ParallelLMHead, DEFAULT_VOCAB_PADDING_SIZE)
+    DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
 from vllm.model_executor.parallel_utils.parallel_state import (
     get_tensor_model_parallel_world_size)
+from vllm.model_executor.sampling_metadata import SamplingMetadata
 from vllm.model_executor.weight_utils import (default_weight_loader,
                                               hf_model_weights_iterator)
 from vllm.sequence import SamplerOutput